Fit a Heterogeneous Survival Effect Forest — fit_survival

Wraps grf::causal_survival_forest() with a column-name interface and computes out-of-bag effect predictions through predict().

Usage

fit_survival_forest(
  data,
  time,
  status,
  treatment,
  covariates,
  target = "RMST",
  horizon = NULL,
  num.trees = 2000,
  seed = 1,
  ...
)

Arguments

data: A data frame containing all required columns.
time: Name of observed time column.
status: Name of event indicator column (1 = event, 0 = censored).
treatment: Name of treatment column.
covariates: Character vector of baseline covariate column names.
target: Prediction target passed to grf::predict().
horizon: Optional horizon passed to grf::predict().
num.trees: Number of trees passed to grf::causal_survival_forest().
seed: Random seed passed to grf::causal_survival_forest().
...: Additional arguments passed to grf::causal_survival_forest().

Value

A heteff_fit object.

Examples

if (FALSE) { # \dontrun{
set.seed(2)
n <- 300
df <- data.frame(
  time = rexp(n, 0.2),
  status = rbinom(n, 1, 0.8),
  trt = rbinom(n, 1, 0.5),
  x1 = rnorm(n),
  x2 = rnorm(n)
)
fit <- fit_survival_forest(
  data = df,
  time = "time",
  status = "status",
  treatment = "trt",
  covariates = c("x1", "x2"),
  target = "RMST",
  horizon = 5
)
head(as.data.frame(fit))
} # }